Bad news for all of you who welcomed a little respite from AI commentary: we’re back, baby! This last quarter has seen the litany of use cases — real and hypothetical — continue its procession through news media. Here are some of our favourites:
- AI being used to decipher the text of a volcano charred Roman scroll from Pompeii.
- AI helping to find the three-quarters of all industrial fishing vessels and a quarter of transport and energy ships that are sailing incognito. (Who knew? We’d assumed we knew where the boats were this whole time.)
- AI generated music — which we’ve actually been using at Novigi for comic reasons. It’s awful.
Two stories struck us as perhaps more relevant to our clients and partners in wealth management:
- Capital investment in AI
- The impact on AI decision making
Capital investment in AI
In March 2024, AustralianSuper’s Chief Technology Officer, Mike Backeberg, announced that the fund would be participating in the Copilot for Microsoft 365 Early Access Program (EAP). AustralianSuper joins Rest Super and Suncorp who announced their participation in the EAP in late 2023.
Backeberg was bullish about the potential for generative AI to transform superannuation, and for AustralianSuper to lead that transformation. “We aim to be thought leaders in the generative AI space,” said Backeberg. “Our core priority is to use it to improve member outcomes, both through the way we provide services and how we are using it to free up our people to focus more on higher-value tasks.” AustralianSuper’s goal is to use AI to yield a 5% increase in productivity within 12 months. Backeberg reckons this is conservative, though he notes that at even 5% the increase in productivity will pay for the technology “multiple times over”.
Juxtaposing this enthusiasm, is reporting on economy-wide capital expenditure and the propagation of AI. The Economist points out that investment booms were required to see the adoption of previous transformational technologies, from the tractor to the personal computer. Using the US economy as a proxy — which is notably rosier than Australia’s at present — we see capex intentions increasing just 2.5% year on year as of 31 March 2024. This is less than the probable rate of inflation over the same period.
If the numbers are to be believed, we are not yet seeing the private sector investment required for AI to truly transform the economy. Backeberg noted that AustralianSuper would “have to think carefully about how we embed it [AI]” into its processes. Indeed, even if the sticker cost of tools like Microsoft’s Copilot are relatively low, making the organisational changes required to embed it and use it efficiently will require significant investment. For the moment, it seems as though firms are not ready to make it.
The impacts of AI on decision making
Behavioural economist David Almog and his coauthors published a paper in preprint earlier this year examining the impact that Hawk-Eye — an AI-based system — has had on decision making in tennis umpiring. While the connection back to wealth management may seem tenuous, we think it’s a safe assumption that generative AI initiatives like AustralianSuper’s will invariably see staff using AI to support decision making.
Almog et al found that with Hawk-Eye having the final say, human umpires made 8% fewer mistakes overall than they did prior to its introduction. This is good news for superannuation funds; a decrease in human error could play an important part in achieving a 5% increase in productivity. Interestingly though, the types of errors that human umpires made changed. Before Hawk-Eye, human umpires were more inclined to make type II errors (e.g. calling a ball out when it was in fact in). With Hawk-Eye, umpires were more likely to make type I errors (e.g. calling a ball in when out).
Superannuation funds implementing AI should take note of these findings. We know that significant investment will be required to ensure that AI is integrated into systems and processes. What Almog’s research makes clear is the importance of understanding the impact of AI on the decisions that staff make. Funds will need to ensure that they have the business intelligence and data management capabilities in place to measure and monitor how AI tools change the behaviour of operational staff. What kinds of biases do they exhibit? What kinds of errors do they make? Understanding these shifts will be essential to managing and mitigating any new risks that may arise.
This article was produced as part of The Quarterly – Data and Technology in Superannuation, Q3 FY24
For more information about anything you’ve read here, or if you have a more general inquiry, please contact us.
Key Contributors:
Kevin Fernandez is General Manager, Market Strategy and Propositions at Novigi, and is based in the Melbourne office.
Sophie Bowen-James is an analyst in the Market Strategy and Propositions team at Novigi, and is based in the Sydney office.